Why blender causing my PC to restart?

@Felix_Kutt Today i went to that shop to test my mobo gpu and psu. But looks like they only have an old 2nd gen mobo, rx 460/550 2GB and no unpacked branded psu more than 450W :confused:
But 450W is not enough for my rig, so at first he changed my mobo and cpu with that i3 2nd gen one.
asus mobo couldn’t detect my windows so he had to reinstall it again :confused:
After that he told me to check if it’s still restarting or not, so i installed substance painter again and tried to export 8k maps, for the first run pc didn’t restart but the software crashed, again i reopened substance painter and tried to export 8k maps again, pc didn’t restart but this time screen went black (Note : he was using e old 720p monitor with HDMI to vga converter, and as far as i can tell they’re no experts)
So after that he told me that maybe my mobo is fine and gpu or psu is faulty, Then he plugged his rx550 maybe or 460 not sure which is only 2GB.
and after installing it i had no 8K export options, only 4K export. So this was a fail test, I was wondering what should i do now, when i was coming out of the shop he told me that he has a 650W psu on his pc in home. if i want he’ll test my pc with that psu on friday.
So what do you think? can you guess the problem now?
And i also contacted the amd support, but they told me to clean install amd driver which i already did, and also told me to update bios, i think which is irrelevant.

The mobo is a MSI X470 Gaming Pro. I don’t play games much but thought that a mobo designed for good graphics would work well with Blender. Could have been wrong though, I’m no expert in these matters.

Yes, Blender was working fine, if somewhat slow. before the upgrade but that was due to the lack of a good CPU.
I have run a couple of stress test programs. One called Stress for the CPU and one called Glmark2 for the GPU. I have encountered no problems/crashes when running them.
I haven’t tried the RAM using one stick at a time yet but have pulled them out of their sockets and plugged them back in, I also did this with the graphics card and gave it a clean. Will try the one stick of RAM at a time later this evening and report back on the results.

Either there is a VRM related issue, could be the one on the motherboard(for CPU) or onboard the GPU. OR it is your PSU, and it might be simply under-powered and I’m not talking so much about the Wattage here.

Every PSU should have some info typically in the following format:


Image from: https://forums.tomshardware.com/faq/power-supply-101-understanding-power-supplies-and-selecting-the-right-one-for-the-job.1608758/
On the image you see a box highlighting the max load stats in Amperes. This could be whats causing this, potentially.

1 Like

Its very simple. Nothing in software is relevant. Do you know what a line filter is?
Wait. You mean the blender is not crashing some other computer? It is crashing itself? In which case all those discussions about electricity (ie amperes) had no relevance whatsoever.

And still, things like Ram have no relevant to what a power controller does.

Felix_Kutt’s reply assumes the power controller is seeing a defect in the PSU - and then shutting off power. That is one reason for an error 41 message and what it is suppose to do. That defect can be identified using a digital meter. Numbers will identify the defect that a controller sees.

Add another one for crashing entire OS.

Windows10 Pro (Bootcamp) on a i7 iMac 5K
AMD RX 580
40GB ram

File is 7GB and when I try to render it crashes Windows and restarts my computer.

Same file on OSX works but it’s not using the GPU obviously. So this maybe a Blender and AMD driver issue with in Windows?

Tried out bootup with one memory stick at a time and no problems to report. Can’t find any other hardware issues and it only when using Blender volumetrics that my PC crashes. Strange stuff.

Blender, as in the blender program that this site BlenderArtists uses is a program, or software.

And yes, I do know what a line filter is, I do have an electrical engineering degree. And while I can see how a program that is pushing the PSU to its limits on a rail could cause a pc to hard boot. I suspect a program giving off EM noise is not the cause of it. And are you sure you want to tell a digital artist to take a digital meter into the powersupply to troubleshoot? That is kinda…not safe.

While I appreciate your recommendations here, we should say very loudly that electricity is dangerous stuff and if you’re to a point where you’re poking around the power-producing parts of your machine with a multimeter, you better have prior experience doing that sort of thing. Otherwise, your better solution is to find a professional.

That said, PC parts are sold as modular units. There’s nothing superior about replacing a sub-component on a motherboard or PSU rather than replacing that whole module.

2 Likes

Update on my restarts: Downloaded today’s 2.81 build and also had an new Windows update waiting (1909) I installed both an now Blender is working as normal.

Not sure which one solved the issue of crashing the OS, but the Blend file at fault will render now.

3 Likes

Here’s mine, what about this one?

Well, not sure who manufactures the PSU’s under Thermaltake* branding, and the fact that they have a 87% efficiency label on there imitating the bronze rating is a little concerning, but if we can trust the actual info on that sticker then 48A or greater and 588W on the 12V rail should be fine.

* Many brands just put their own sticker on a PSU that’s actually manufactured by another company. Some of the best manufacturers that license their PSUs for other brands like that are SeaSonic and SuperFlower, but there’s a myriad of no-name manufacturers of more questionable quality as well. Corsairs top end is typically manufactured by SeaSonic while EVGA usually uses SuperFlower AFAIK. So Corsair and EVGA are a couple of fabless brands you can trust. And SeaSonic and SuperFlower and couble of Fabs you can trust.

For now though I’d recommend giving @CookItOff solution a try.

Most all computers consume less than 350 watts. But most who build computers have no idea how to select a PSU. So we tell them to install a 700 watt supply. Then help lines are not clogged teaching high school science concepts to computer assemblers.

The disk drive typically consumes less than 2.5 watts. So those programs that recommend power supplies say a disk drive consumers 10 or more watts. Since those target computer assemblers,.

A 700 watt supply should not cause problems detected by the power controller - not matter what the load. Even a 450 watt supply (selected using other knowledge) is more than oversized. Watts really say nothing accurate. As the red circle on that picture demonstrates, only amps for each voltage are relevant.

Using a digital meter can identify many power problems that would be identified by that power controller. Which one is problematic? Nobody can say without numbers from that meter. Even voltage monitoring software cannot report all necessary numbers. (And a volt meter on the motherboard must be calibrated.)

Error 41 says a power controller has observed a hardware defect that is so bad as to shut off power. The defect exists constantly. Defects and failures do not always coincide. That meter (used with proper instructions) can often see a defect even when that application program is not causing an error 41.

Forget the AC line filter. It is now irrelevant.

Nobody can accurately say what is sufficient for each rail. Since it can vary for different hardware. And so much misinformation is posted (ie disk drive consuming more than 10 watts, PSUs that really do not output listed current due to various reasons). Only way to confirm those amp numbers are sufficien, for that unique PSU and other hardware,t is actual measurements.

1 Like

You do know that Blender 3d is a program, right? But I’m still interested in knowing how you would attach an electrical filter to software.

If we assume that the PSU is the issue then it’s likely the current protection that is triggering on the PSU end(to protect itself and the system). Otherwise it’s likely some VRM related. Only the OP can really test for and determine which part it might be, if it really is HW.

Beyond that I don’t really have much more to offer right now. In my last post I recommended OP tries @CookItOff solution from post #150:

If that doesn’t work, then back to the HW suspicion.

@Felix_Kutt Need urgent suggestion, i called that shop again and they said they only have a antec 500W psu to test on my pc, the other day he used a 450W psu, but for my pc minimum psu req is 500-550W.
So should i give it a try with 500W psu? can this 500W psu able to handle my rx 590?

Antec is one of those brands that slaps their own name on other manufacturers PSU’s. I can’t comment much on it. If you find out the specific model number we may be able to look up the actual make of the PSU perhaps, but I honestly can’t promise much.

Also, did you try @CookItOff 's advice, and did you try the under clocking & volting? Anything to report about that?

edit: ok, so this is probably it(but again just my blind guess, without model number I can’t determine much):
https://antec.com/product/power/vp500p-plus.php

Looking at the stats they feature about it, it looks like it would be worse than your original/previous PSU at least.

looking at the following topic on toms hardware and considering that the one replying is the legendary jonnyguru himself:

If it is the exact same psu, then to quote him at least “it’s a lot better than a lot of the no-name garbage I see sold in India.”

Still, he only says it’s ok for light use. Stressing your CPU and GPU with blender is most definitely not light use.

@Felix_Kutt Nvm that,
i just went to that shop, since that antec psu wan’t enough or has some issue but pc was restart randomly.
But luckily someone came with is pc he was facing some other problem, but atleast for him i could try his gpu (rx 580) and his corsair vs 650W PSU, and still the pc was restarting when i tried to export 8K textures wtf.
Ram,Gpu,Psu,SSD,HDD we tested everything, also tested with another motherboard, but i think mobo test wasn’t a success. in that mobo first software crashed 2nd. Display went black.
So what is the problem now? don’t think cpu can’t cause this issue unless it’s overheating, but in this case cpu isn’t overheating.
So where is the problem?

Nothing in a manufacturer name or any other subjective reasoning will answer that question or solve the problem. Worse, having made so many changes now means an additional problem can exist.

That answer could have been obtained in minutes using a meter and some requested instructions to perform two minutes of labor. A task that even a 12 year old can do. Resulting three digit numbers then means another, who actually knows how hardware works, can say what is suspect and what is exonerated. Without any expressions that say, "maybe it is … ", “try that.” , or swap parts.

Some fundamental concepts apply. For example an assembly of all good parts can still fail. Only useful conclusion comes from viewing what all those parts do together. And that means numbers.

How to solve this problem is clearly stated in an above sentence. That means the defect is first defined long before even one part or cable is removed or disconnected.

Bro, i have experience the same problem, i have been struggling with it for a week and cannot even using PC.
I thought it was software problems. but after checking and reinstall window for 3 times, the problem still there. so i thought maybe the hardware had some problems. i test every single unit but none of them had problem. i disassemble and reassemble the whole PC 4 times, nothing works.
And finally, i bought a air duster and cleaning every slot on the mother board. then it works.
I am not sure what exact problem it is, but please spend 6 bucks buy a air duster and carefully cleaning all the hardware totally. it maybe some small dust cause the problem.
Good luck

1 Like